How it works
InfoQast gives every client of your agency a working AI team — with you as editor-in-chief. Here's the whole pipeline, from brief to report.
1. Brief goes to the Strategist
You describe the client's month — goals, campaigns, constraints. The Strategist turns it into a content plan: topics, formats, and cadence for each connected network. It works from the brand profile and from what the Analyst learned last month, not from a blank prompt.
2. One master post, compiled for every network
The Copywriter writes a structured master version of each post; the Designer produces visuals in Creative Studio within your brand kit. Then the compiler turns the master into native variants — a thread here, a carousel there, a caption with the right hashtags and length limits everywhere. You can edit any variant; protected fragments survive recompilation.
This is the part that eats most of an agency's week — and it's exactly the part that becomes minutes.
3. Your approval gate
Every brand runs at an autonomy level you choose. The default is conservative: AI proposes, you publish.
L0Suggest
Agents draft; every variant waits for your explicit approval before anything is scheduled.
L1Prepare
Agents draft and schedule into a queue that still requires approval to release.
L2Publish routine
Pre-approved content types go out on schedule; anything new or sensitive still stops at the gate.
L3Autopilot
Full-cycle publishing for a brand — an explicit opt-in, guarded by preflight checks, moderation, and a kill switch.
Clients can approve too: the client portal gives each customer a clean view of their drafts and an approve button — without seeing your other clients or your costs.
4. Publish via official APIs, report with real numbers
Approved variants go out on schedule through official platform APIs — today that's Telegram, Bluesky, Mastodon, Discord, and YouTube, with Meta, TikTok, and LinkedIn connecting through reviewed apps next. The Analyst collects per-client metrics and turns them into a report your client actually reads — branded as your agency on the Agency plan.
Your methodology becomes an asset
Brand voices, hooks, content structures, taboos — everything your agency knows lives as skills: versioned files the agents actually use. Review them, edit them, roll them back. When the AI team learns from a month's metrics, the learning lands in the skill — not in an employee's head that can resign.
Built for trust
Preflight checks
Every variant is validated against each network's rules and your brand policy before it can be scheduled — problems surface at the draft stage, not at publish time.
AI moderation layer
A separate compliance pass reviews content before delivery; a kill switch stops all publishing for a brand or a workspace instantly.
Audit trail
Every agent action, approval, and publish is logged. You can always answer "who did what, and why" — to yourself or to your client.
Official APIs only
We publish through each platform's official API and never automate personal profiles or work around limits. Your clients' accounts are not an experiment.
And the money stays visible: every agent action is metered per client at official Anthropic list prices, so you always know what a client costs before you quote them.
Take the next client without the next hire
The alpha is invite-only: we review requests weekly and onboard agencies in cohorts of five. Design partners get locked pricing and a direct line to the founders.